The signal processing libraryis missingippsCompare
(<, >, =. >=, <=) routines. They can get very handy in
some cases. Together with them you would have to
provide "gather" and "scater" methods to allow
compressing and decompressing sparse data
returned from ippsCompare (sparse vector
support). For f32 and f64 types. Complex
support would be absolutely great.